摘要
提出了一种基于LCS的特征树最大相似性匹配网页去噪算法。通过将目标网页和相似网页转化为特征树,并将特征树映射为一个特征节点序列,利用LCS算法能获得最长子序列全局最优解的特点,找出两棵特征树之间的不同节点作为候选集,并对候选集进行聚集评分找出网页重要内容块。给出了算法的原型系统,并对每一个模块的实现做了详尽的描述。
A maximum similarity matching algorithm for noise reduction in Web pages is presented based on LCS. Parsing target page and similar pages into two characteristic trees, and mapping them to two characteristic node sequences, the LCS algorithm can get the longest sub-sequence which is global optimal solution, and the different characteristic nodes is found out between the two characteristic tree as a candidate set, clustering the candidate set and scoring to identify web page important informative block. In this paper, the algorithm prototype is given, and the implementation of each module is described.
出处
《电视技术》
北大核心
2011年第13期44-48,63,共6页
Video Engineering
基金
国家"863"计划项目(2008BAH28B04)
上海市科委资助项目(08dz1500108)
中国博士后科学基金
上海市博士后基金资助项目(20090460637
10R21414800)
关键词
LCS
特征树
网页去噪
LCS
characteristic tree
noise reduction in Web pages
